RETRIEVAL OF OPTICAL CONSTANTS AT UV-VIS-NIR FOR MARTIAN
DUST ANALOGUES BY MODELLING LIGHT SCATTERING.
J. Martikainen, J. C. Gómez Martı́n, O. Muñoz, Instituto de Astrofı́sica de Andalucı́a, CSIC, Granada, Spain
(juliamar@iaa.es), T. Jardiel, M. Peiteado, Instituto de Cerámica y Vidrio, CSIC, Madrid, Spain, M. Wolff, Space
Science Institute, Boulder, Colorado, USA, and the RoadMap team.

Introduction
Particle composition in light-scattering and radiativetransfer models is parameterized through the wavelengthdependent optical constants, i.e., the complex refractive
indices (m = n + ik). The real part of the complex refractive index, n, describes the ratio of the speed of light
in a vacuum to the phase velocity of light in the material,
whereas the imaginary part, k, describes the absorption
of light inside the material.
Optical constants are needed from modelling singleparticle scattering within the atmospheric dust to simulating the global Martian climate. Reliable complex
refractive indices of Martian dust are difficult to find in
the literature for the wavelengths used in the simulations.
Previous studies by Wolff et al. (2009) & (2010) have
obtained optical constants from analysing the observed
dust storm spectra at 258-2900 nm region by using cylindrical particle shapes, however, these values could not
be supported with direct laboratory measurements.
In this work, we retrieve the complex refractive indices of three Martian dust analogues at UV-vis-NIR
wavelengths by using the measured size distributions,
diffuse reflectance spectra, and an advanced light-scattering
model with realistic particle shapes. We compare the retrieved values to those in literature.

Figure 1: The Martian dust samples. Top row is the JSC
Mars-1 sample with 63-100 µm (left) and 20-40 µm (right)
size fractions, middle row is the MMS-2 sample with 63-100
µm (left) and 20-40 µm (right) size fractions, and bottom row
is the MGS-1 sample with a 20-40 µm size fraction.

Martı́n et al. (2020). In Fig. 2, we show the measured
projected surface area distributions S(r) and number
distributions n(r).

Samples and Size distributions
For this study, we selected three Martian analogues relevant to the atmospheric dust of Mars: JSC Mars-1
(Johnson Space Center regolith simulant), MMS-2 (Mojave Mars Simulant), and MGS-1 (Mars Global Simulant). Each sample represents a different region on Mars
and has a distinctive chemical composition, texture, and
color (see Fig. 1).
The JSC Mars-1 and MMS-2 simulants were processed to produce two well-defined narrow size distributions in the geometric optics region (r >> λ): a particle
size fraction from 63 to 100 µm, and a second fraction
from 20 to 40 µm in diameter. The MGS-1 simulant was
processed until a particle size fraction from 20 to 40 µm
was obtained.
We measured the size distributions of the final samples with a laser light scattering particle sizer (Malvern
Mastersizer 2000). A detailed description of the particle
sizer and the retrieval process can be found in Gómez

Spectral Measurements
The diffuse reflectance measurements were carried out
over the range of 200 to 2000 nm using the Cary 5000
UV-vis-NIR integrating-sphere spectrophotometer at Centro de Instrumentación Cientı́fica, University of Granada.
The instrument is equipped with a deuterium arc source
for the UV wavelengths and a tungsten-halogen source
for the vis-NIR region. Each measurement was calibrated using a polytetrafluoroethylene (PTFE) reflectance
standard. We present the measured spectrum of the
MGS-1 simulant in Fig. 3.
Model
To retrieve optical constants from a reflectance spectrum, a light-scattering model is needed for the sample

particles. We derived the imaginary parts of the complex
refractive indices, k, using a model based on the retrieval
by Martikainen et al. (2018). The code was upgraded to
take into account different size distributions. We fixed
the real part of the complex refractive index to 1.5 as it
does not change significantly at the modeled wavelength
region.

Figure 4: Example particle shapes used in the model (right)
compared to the JSC Mars-1 63-100 µm size distribution FESEM image (left).

Figure 2: Projected surface area distributions (top) and number distributions (bottom) of the JSC Mars-1 20-40 µm and
63-100 µm fractions, the MMS-2 20-40 µm and 63-100 µm
fractions, and the MGS-1 20-40 µm fraction.

Figure 3: The measured reflectance spectrum of the MGS-1
sample.

The retrieval was carried out with an optimization
code that uses the measured absolute reflectance and the
SIRIS4 (Muinonen et al. 2009; Lindqvist et al. 2018;
Martikainen et al. 2018) ray tracer that simulates light
scattering by Gaussian-random-sphere (GRS) particles
larger than the wavelength of the incident light. The
shapes of the GRS particles were defined by the autocorrelation function of the logarithmic radial distance,
νG , and the standard deviation of the particle radius, σG
(Muinonen et al. 2009). In our model, we used νG = 3
and σG = 0.2 so that the model particles were as realistic
as possible (see Fig. 4).

Figure 5: The derived k of the MGS-1 sample compared to
the values retrieved by Wolff et al. (2009) & (2010).

Results and Discussion
The derived imaginary parts of the refractive indices for
the MGS-1 sample are shown in Fig. 5 together with
the values retrieved by Wolff et al. (2009) & (2010) that
are widely used in the Martian atmospheric studies. The
preliminary results indicate that the values retrieved with
the advanced light-scattering model are much smaller
than those obtained in previous studies. The differences
can be partly explained by the choice of particle shapes:
real dust particles are known to be irregular and thus
using cylinders or spheres will introduce an error in the
retrieved optical constants. Furthermore, the exact grain
sizes in the observed dust storms are poorly known and
may contain material different from the analogues that
we used.
As the three samples used in this work are significantly distinct from each other, the retrieval gives a good
estimation of the range of optical constants of Martian
dust. In future studies, it will be important to use realistic
values to accurately interpret laboratory measurements
and in-situ observations.
Acknowledgements
This project has received funding from the European
Union’s Horizon 2020 research and innovation program
under grant agreement No 101004052 (RoadMap).

References
Gómez Martı́n et al. (2020). Computational study of
the sensitivity of laser light scattering particle sizing to
refractive index and irregularity, Journal of Quantitative
Spectroscopy & Radiative Transfer, 241,106745.
Lindqvist et al. (2018). Ray optics for absorbing particles with application to ice crystals at near-infrared
wavelengths. Journal of Quantitative Spectroscopy & Radiative Transfer, 217,329–337.
Martikainen et al. (2018). Spectral modeling of meteorites at UV-vis-NIR wavelengths , Journal of Quantitative
Spectroscopy & Radiative Transfer, 204,144–151.
Muinonen et al. (2009). Light scattering by Gaussian
particles with internal inclusions and roughened surfaces
using ray optics. Journal of Quantitative Spectroscopy &
Radiative Transfer, 110(14),1628–1639.
Wolff et al. (2009). Wavelength dependence of dust
aerosol single scattering albedo as observed by the Compact Reconnaissance Imaging Spectrometer. JGR Planets, 114, E00D04.
Wolff et al. (2010). Ultraviolet dust aerosol properties
as observed by MARCI. Icarus, 208, 143–155.